The Predicate Structure to Represent the Intention for Software

Transcription

1 The Predicate tructure to Represent the Intention for oftware Fumio Negoro The Institute of Computer Based oftware Methodology and Technology Abstract The requirement in this study is a sentence written in a natural language. This is a methodology to determine programs by mechanically extracting words from the requirement, where the words are nouns in the sentence. The program is called cenario Function (F). F is a set of propositions. The proposition is determined with the predicate structure that takes a noun as its variable. The predicate structure has been derived from this study. The nouns used as variables are clues to realize on F the meaning and the functionality contained in the sentence. The requirement that cannot be expressed only with the nouns is realized by the complementary action of the propositions, which is realized when F is executed on computer. In other words, the most of the work required in the conventional ways of software development such as requirement definition, logical design, program design and tests, can be eliminated and is no longer the primary part of software development with this methodology. As a result, a human work in software development can be changed into a work with less burdens.. The Core of non-engineering hypothesis In this section, the core of the hypothesis is briefly mentioned. We have studies what an intention is. The conclusion is that an intention is made when a correspondence between two worlds is made and that the two worlds are materialized with a set of particular elements. Time has a variety of velocities that are independent elements. The elements are nonexistent to us. These elements are the elements of the two worlds. The separation of the worlds is made based on the difference of the time velocity. One of the worlds is called the connotational world and the other is the denotational world. If a correspondence is made between the two worlds, that state becomes an intention. At that moment, the elements in the connotational world are objectified. The objectified element cannot return to the state that it has been objectified. The behavior of the objectified elements is the whole of the object of our cognition. However, we are unable to cognize the whole. entences, requirements and natural languages consist of the objects; therefore, they are never complete as they cannot represent the whole. The model to realize this hypothesis is the Consciousness Model, where the above hypothesis is made into axioms. []. TDM and F According to the axioms, a requirement is a set of objects reflecting the intention. Therefore, it is not that we are able determine a requirement autonomously. All depends upon the intention. The purpose of this study is to find a language to represent the intention. However, in this paper, a methodology to represent the intention in a programming language is presented. The scheme is to be realized, which simultaneously satisfies a process of reversing a requirement that comes to someone s mind at a certain moment to the world where the intention has been made, and a process of enabling the creation of a requirement that fully reflects the intention. A structure is to be realized, which leaves the above processes to an action of computer, manifests a requirement into the memory of computer, and makes the intention in the mind of the computer users. The scheme is defined as Three- Dimension-like pace Model (TDM) by the axiomism, and it is expressed in a programming language so as to leave it to a computer. This is the cenario Function (F). [] [] TDM is a static state of the aforementioned scheme, and F is the same. F changes a static state of itself into a dynamic structure on computer. TDM and F on execution realize the hypothesis of the Consciousness Model. The two worlds are placed in the space in TDM, and nouns are placed on the three coordinates of TDM. The rules of placement of the nouns are determined according to the nature of each coordinate. The nature of the coordinate is derived from the axioms. There exist the particular elements in the two separate spaces in the TDM. Those elements are not cognizable.

2 However, a certain correspondence is made between the nouns on the coordinates and the elements. The correspondence is also defined by the axioms. This correspondence is the predicate structure. The predicate structure is a structure to redefine the nouns as the hypothesized elements. When this predicate structure becomes propositions, they are called ignification Vectors. The origin of TDM is determined when the intention is made and objectification of elements is done. In other words, the intention means a closed boundary line of a certain domain according to our definition, and the inside of the domain corresponds to the connotational world and the outside corresponds to the denotational world. A set of the origins of TDM corresponds to the boundary line. Therefore, execution of F simultaneously defines the origins and a set of them. The meaning of execution of F can be explained in another way. If the number of object of someone s requirement is N, F is determined by N, whereas when F is executed, N N objects are created by the complementary action as a possible number of the objects. In the conventional ways, only N objects can be turned into programs and the N object must be determined logically beforehand. On the contrary, with this methodology, N ignification Vectors are determined deterministically and N N objects are realized.. Definition of F F is a structure to realize TDM by using computer. F can be defined based on a set of nouns that manifest a certain psychological state. A screen and a file are a set of nouns. They become a basis to define F. To develop software concludes to define F. In the following, a general structure of F is defined. F a =Φ[Φ({L(j)}+{O(r)}+{(a)}+R) +Φ({L(i)}+{I(r)}+R) +Φ({L(j)}+ {RR, RD, RM, RC})] a Herein, {L(j)}, {O(r) }, {(a)}, R, {L(i)}, {I(r)}, R, {L(j)} and {RR, RD, RM, RC} are generally called Tense Control Vector. There are kinds in all. L(j), L(i) and L(j) are called ignification Vectors, whereas the rest of the Tense Control Vectors are generally called Action Vectors. O(r), I(r) and (a) are Output Vector, Input Vector and tructural Vector respectively. R, R, RR, RD, RM and RC are Routing Vectors. The definition and role of each Tense Control Vector is explained below. The ignification Vector is defined for every noun. The j and i respectively represent output and input attribute of a noun. The role of L(i) is to duplicate Programs of the F are shown as Lyee cenario Function sample programs at information from an input memory area to its own memory area. The role of L(j) is to create information to be stored in its own memory area by complementary action. The role of L(j) is to store in its memory area an execute directive of ignification Vector L(j). The O(r) is defined for every output memory area, r. The role is to output data in the output memory area to an output physical unit. The output attribute noun, j, is a noun belonging to the output memory area. The I(r) is defined for every input memory area, r. The role is to obtain data from an input physical unit into its input memory area. The input attribute noun, i, is a noun belonging to the input memory area. The (a) is defined for every memory area, a, requiring initialization. The role is to initialize the area. R, R, RR, RD, RM and RC specify Pallet Function that is executed next. The Tense Control Vectors are bundled by the three kinds of Pallet Functions, Φ, Φ and Φ, as follows: Φ({L(j)}+{O(r)}+{(a)}+R), Φ({L(i)}+{I(r)} +R), and Φ({L(j)}+{RR, RD, RM, RC}. The symbol {} is a set symbol to bundle Tense Control Vectors, and the + is a symbol to indicate the execution order. For example, Tense Control Vectors bundled in a Pallet Function, Φ, are placed in the order of L(j), O(r), (a) and R, and they operate in this placement order. This is expressed by Φ({L(j)}+{O(r)} +{(a)}+r). In a set of kind-wise Tense Control Vectors, a symbol to indicate execution order is not contained. For example, the set L(j) means that the way of the placement of the ignification Vectors of L(j) is free, and it operates in the free-placement order upon execution. A set of Tense Control Vectors bundled by the Pallet Function is generally called Pallet. The Pallets to be bundled by Φ, Φ and Φ are respectively expressed as W0, W0 and W0. The Tense Control Function, Φ, assigns executionright to the next-to-operate Pallet Function. A Pallet Function having been given the execution-right gives the execution-right to all Tense Control Vectors belonging to the self. Once Tense Control Vectors are thoroughly executed, the execution-right is returned to the Pallet Function. The Pallet Function judges whether its Pallet needs to be re-executed or not. If the re-execution must be repeated, the execution-right is again given to all Tense Control Vectors belonging to the self. If otherwise, the execution-right is returned to the Tense Control Function. Thereafter, the Pallet Functions make execution in the order of Φ, Φ, Φ, under the Tense Control Function. It is defined in the format of Φ(Φ+Φ+Φ). This is an expression of the F definition. The execution of F and the Tense Control Vectors are described the sections and.

3 . Predicate structure If an element of the denotational world that corresponds to an element of the connotational world exists, the element is made exteriorized because it can actualize as a real existence with view to the hypothesis. However, if it has already been actualized, it is not necessary to actualize in repetition. Also, if there is no corresponding element in the connotational world, an element of the denotational world cannot be actualized. The actualization means to make the self actualize by utilizing other element that has already been made into a real existence. This is a complementary action. However, the complementary action must be what reflects an intention. It is implemented in accordance with the providence of actualization. This means a state in which the mutual relation concerned with the complementary action establishes ynchronous tructure. Through this, the actualization is recognized. If otherwise, the complementary action must be done over again, or, stop doing it. All mentioned above is represented as the predicate structure shown in Fig.. The predicate structure consists of seven boxes and each box contains a specific rule. The definition of all operators that are shown in the definition of F is effected by the determination of the seven kinds of rules of the predicate structure. The st rule is contained in the first box, and so on. ach of the nd, th, th and th boxes has a memory area where the result of execution of the rule is to be stored. Those memory areas are called the nd, th, th and th areas respectively. The predicate structure becomes a ignification Vector when a noun is given to it as its variable, whereas the predicate structure becomes an Action Vector when it takes a plural number of nouns as its variable and performs a specific role. Figure. Predicate structure The six kinds of rules except the nd rule are autonomously determined based on the definition information of the nd rule, so the definition action is a work for the nd rule only practically. The nd rule is already fixed by what must be regulated by each operator, so F can be defined simply by implementing the regulated work. This work is already done with a tool. Associated work elements required to execute F on computer, for example, logical and physical definitions of a screen must be fulfilled separately from the above work as a natural course of procedure. A ignification Vector is a proposition because it is defined by taking one piece of noun as its variable, and an Action Vector is a function which is defined as a functionality that takes plural nouns as variable. The ignification Vector is true as a proposition if a value is being set in its th area. The complementary action of ignification Vectors materializes as a relation of propositions that become true. The Input Vector materializes as a function if information can be obtained in its th area from an input physical unit. The Output Vector materializes as a function if a flag indicating success of an output action is set into its th area. The Routing Vector materializes as a function if the next-to-execute Pallet is appointed in its th area.. ynchronous structure In this study, a relation to establish an intention and a requirement of its reflection concurrently is called ynchronous tructure. That is, it is a world of TDM and F. The synchronous range means the range of an intention to materialize as well as the range of a requirement of its reflection. A plural number of Fs can be connected each other. In liaison with one another upon execution, plural Fs realize expansion or reduction of the synchronous range. ame as F can be defined statically, a state providing the liaison can also be defined statically by using Fs as elements. The diagram to show the liaison of Fs is called Process Route Diagram, hereinafter denoted as PRD. F can be defined as a set of nouns which are unitized as a requirement, for example, by taking one screen and a set of nouns belonging thereto as a unit. Therefore, a F can be defined in correspondence to each of the screens in the screen transition. By taking Fs as elements and based on the relation of the screen transition, PRD can be defined. Owing to the F s structural features, PRD can be drawn through mechanical procedure if only the information of individual screens can be defined. Therefore, if this procedure is mechanized by a tool, a work to define the screen transition information can be substituted by the tool. A rule of liaison of plural pieces of F is four kinds. It is reflected as a definition of Routing Vectors, RR, RD, RM and RC, belonging to W0.

4 . xecution of F Once a Pallet s Tense Control Vectors are executed thoroughly, the status of the Pallet is recorded. Recorded is a status of just before and current. The status of a Pallet means a set of status of the th area (true/false as a proposition), the th area (refusal flag) and the th area (restart flag) of all the Tense Control Vectors belonging to the Pallet. These areas are generally called Pallet Area in particular. Routing Vectors belonging to W0 and W0, i.e. R and R, compare the just-before and current status of the self Pallet, and if there is a state transition, the execution of the self is terminated. If there is no state transition, the Pallet to be executed is appointed in the self s th area. R appoints W0 of the same F, R appoints W0 of the same F. Pallet Functions of W0 and W0, i.e. Φ and Φ, implement check on the th area of R and R belonging to the self Pallet, once all Tense Control Vectors belonging to the self Pallet are executed thoroughly. Then, if the next-to-execute Pallet has been appointed thereabouts, the Pallet Function transfers executionright to the Tense Control Function, Φ. Based on the information, the Tense Control Function transfers execution-right to the next-to-execute Pallet Function. If the next-to-execute Pallet has not been appointed thereabouts, the Pallet Function gives execution-right to the Tense Control Vectors belonging to the self and reexecutes from the beginning. There are four kinds of Pallets to be appointed as the next-to-execute Pallet by the Routing Vector belonging to W0. For this reason, four kinds of Routing Vectors, RR, RD, RM and RC, are defined on W0. Details of Routing Vectors are described later. The status of Pallets symbolizing the appearance of the F execution is a result of executing ignification Vectors. It is a result of the complementary action to materialize among ignification Vectors to be executed. Therefore, the presence of the state transition of Pallet Areas is a condition to determine whether F is to be executed or not. That is, if a state transition occurs, it indicates that there is still a possibility of materializing complementary action in the F, and if no state transition, it indicates that a complementary action does not materialize. Based on the data which is obtained with the Input Vector, F establishes the complementary action autonomously among W0 ignification Vectors. Then the Output Vector determines the state of the result of the complementary action so as to establish another complementary action with us. That is, the Output Vector performs a role to manifest (output) a result of the complementary action established by F. It is synonymous with the materialization of a new intention in our mind. In other words, all the possible requirements of N N materialize by the F s complementary action, and an intention materializes in us by the Output Vector s action. Then, an additional requirement materializes in us, reflecting its intention, thereby an additional complementary action being established in F. The structure to realize the above is the ynchronous tructure. The three-dimensional space of TDM represents our mind, whereas the threedimensional coordinates represents N N requirements. F and we share the N N requirements. In other words, the Input Vector commences to build the ynchronous tructure, and the Output Vector destroys it. Thus, the hypothesis is realized, which an intention materializes in us and then the N N requirement reflecting it materialize. The traditional program establishes logical action before executing on computer. In contrast, F establishes it upon execution through the complementary action that is implemented autonomously by utilizing the state transition of Pallets, which appears structurally. As a result of execution, F produces the same result as the traditional program, but it materializes a meaning fundamentally different from the meaning of movements of the traditional program.. tructure of tense control vector In this section, the rules of each box of the predicate structure as the ignification Vector and the Action Vector are described. The relation of memory areas of the ignification Vector and the Action Vector is shown in Fig.. i i RAD L(i) I(r) Figure. Relation of memory areas j j WRIT L(j) O(r)

5 .. ignification vector The ignification Vector executes the self by itself by obtaining the execution-right from the Pallet Function. Upon the end of execution, the executionright is returned to the Pallet Function. The ignification Vectors on W0, W0 and W0 have different roles as described below.... W0 signification vector, L(j). The st rule compares and checks the presence of the data in the th area of itself and a value in the th area of W0 ignification Vector of the same noun. If a self data exists in the th area of the W0 ignification Vector, the purpose of this ignification Vector was accomplished, so execution-right is returned to the Pallet Function. A set of the th areas of L(j) corresponds to the nd area of the Output Vector, O(r), which is the memory area for output. If there is no data in the th area of itself and if the data in the th area of W0 corresponds to the st rule of W0, the nd rule is executed. The nd rule is to operate an arithmetic calculation or duplicate the information existing in a specific memory area. This is the simplest requirement of users. In other words, the ignification Vectors can be defined by requirements on this level. The nd area of L(j) is a temporary memory area of this ignification Vector. The rd rule checks the presence of self data in the nd area of O(r). If there is data in that place, the validity of the data shall be guaranteed by its presence. The reason is that the self in the nd area of O(r) cannot materialize unless data of other nouns are present in the th areas of L(j) or L(i), which are elements to materialize the complementary action to produce the data in the nd area of the O(r). The th rule of L(j) duplicates this nd area data to its th area, which is the nd area of O(r), then the execution-right is returned to the Pallet Function. If there is no data in its nd area of L(j), the th rule checks the state transition of the Pallet, which is a set of status in the th areas on this Pallet. If there is a state transition, the th rule is executed. If there is no state transition, the th rule is executed. After either of the th rule or the th rule is executed, the execution-right is returned to the Pallet Function. The th rule sets a refusal flag which refuses re-execution of the ignification Vector into the its th area. It causes the issuance of a message if necessary. The th rule sets a recursion flag into its th area, which requests reexecution of the complementary action. The usage of this flag is not fixed in particular.... W0 signification vector, L(i). The st rule checks the presence of data in its th area. If there is data, the purpose of this L(i) has already accomplished, the execution-right is returned to the Pallet Function. If there is no data, the nd rule is executed. The nd rule sets data into its nd area after obtaining it from the input memory area, which is the nd area of the Input Vector. The nd area is a temporary memory area of this ignification Vector. The rd rule checks the presence of data in its nd area. If there is data, the th rule duplicates the data in the nd area into its th area, and the execution-right is returned to the Pallet Function. The th area of L(i) is used by the complementary action of L(j) and the nd rule of the W0 ignification Vector. If there is no data, the th rule is executed. The th rule, the th rule and the th rule perform in the same way as those of the W0 ignification Vector.... W0 signification vector, L(j). The st rule checks the presence of data in the th area of this L(j), and it makes judgment on the necessity of executing itself. If there is data in the th area, the purpose of this ignification Vector has already accomplished, the execution-right is returned to the Pallet Function. If there is no data in that place, its nd rule is executed. The nd rule is to specify necessary condition of the materialization of L(j) of the same noun. The requirement to determine this nd rule of F requires logicality most strongly if the program is developed in the traditional ways. That is, this is the part requiring high-level work knowledge and a corresponding logical design. F has its features at the point of simplifying this problem. That is, this logicality is resolved into the relation of data that establishes the way of mutual relation of plural nouns and their connection. In the traditional world, this must be replaced by a problem of establishing the whole by means of combining both of the sequential order that establishes parts of requirement and the sequential order of those parts. In the scheme of F, each noun is replaced ignification Vectors belonging to W0 and W0, and data are determined for every noun through the complementary action which is implemented autonomously thereabouts. The complementary action is an action to manage the role for this sequential order. The meaning of its autonomous materialization is that the work to determine the sequential order can be removed from human work. The logical condition upon execution is built up resultantly by data for the respective nouns. Therefore, it is clear that this relation defined by the nd rule of L(j) can be handled by the definition of a noun that affects the noun to become a variable. From the relation defined by the nd rules of L(j) and L(i), a directory of nouns can be obtained. From the directory, nouns concerned with the nd rule of L(j) can be obtained. If users define the way of

6 connection of the nouns as a requirement, instructions defined by this nd rule can be conclusively simplified much more than those defined logically by the traditional approach. In this connection, automatic algorithm is established, which converts traditional programs into F. By using it, conversions from Assembler to COBOL and from COBOL to COBOL were implemented. The rd rule checks the validity of data in the nd area, which is a temporary memory area for this ignification Vector. It is enough if the presence of data in the nd area is checked. If there is data in the nd area, the validity of data is guaranteed because of it. If valid, the th rule duplicates the nd area data into its th area, and the execution-right is returned to the Pallet Function. The th rule, the th rule and the th rules are treated in the same way as those of L(j)... Action vectors The mode of giving and receiving of the executionright is the same as the ignification Vectors. The rule of each box of the four kinds of Action Vectors is explained below.... Input vector, I(r). The Input Vector can be defined if logical definitions of a set of nouns and DBM are determined. The input memory area is placed in the nd area of I(r). The st rule makes judgment if input data can be obtained into the nd area. That is, if the input memory area is in the initial state, and if a refusal flag of the {I(r)} is off, data can be obtained, so the nd rule is executed. If not so, the execution-right is returned to the Pallet Function. The nd rule executes an input command such as RAD. If this command is executed normally, it means that data has been obtained into the nd area. The rd rule makes judgment of the validity of the execution result of the input command. This means to examine the presence of data in the nd area, which is its input memory area. If there is data, it means that the execution of the input command has been implemented, the th rule sets a normal flag into the th area. After this, the executionright is returned to the Pallet Function. If there is no data, the th rule is executed. The th rule, the th rule and the th rule are treated in the same way as that of L(j).... Output vector, O(r). The Output Vector can be defined if logical definitions of a set of nouns and DBM are determined, same as the Input Vector. The output memory area is placed in the nd area. The st rule makes judgment if data in the nd area is to be output or not. That is, if a refusal flag of {L(j)} is off, data in the output memory area, which is a set of the th areas of L(j), can be output, so the nd rule is executed. If not so, the execution-right is returned to the Pallet Function. The nd rule executes an output command such as WRIT. If this command is executed normally, it means that data has been output into the nd area. The rd rule makes judgment of the validity of the execution result of the output command. This is defined in accordance with the rule of, for example, DBM. The th rule initializes its nd area and sets a normal flag in its th area if the result of executing the output command is valid. Then, the execution-right is returned to the Pallet Function. If the execution result is abnormal, the th rule is executed. The th rule, the th rule and the th rule are treated in the same way as that of L(j).... tructural vector, (a). The initialization of memory areas becomes indispensable so as to realize re-execution by the Pallet Function. The tructural Vector does area initialization required for it. However, the initialization of the W0 s output memory area is implemented by the Output Vector. A refusal flag and a recursion flag are not treated as an object of initialization. The initialization of these flags is implemented by the Pallet Function. The tructural Vector does initialization of other areas except above. If output success is declared by setting a normal flag in the th area of the Output Vector, it means that the output memory area has been initialized. Following the initialization of the th area of O(r), memory areas of W0 and W0 are initialized. Then, the Input Vector can start to operate.... Routing vectors, R, R, RR, RD, RM and RC. The Routing Vector specifies the next-to-execute Pallet in its th area. Which Pallet to specify is defined by the nd rule of the Routing Vector by using the state transition of the th area of {L(j)} and a refusal flag of {L(j)}. There are six kinds of Routing Vectors and each of which is explained below. R is the Routing Vector of the W0 Pallet. The st rule of R checks the state transition of the W0 to which the R belongs. If there is no state transition, the R s th rule appoints W0 of the same F into the R s th area as a Pallet to be executed next, and the execution-right is returned to the Pallet Function. If there is any state transition, and the execution-right is returned to the Pallet Function. If the next-to-execute Pallet is appointed, the execution-right is returned to the Tense Control Function from the Pallet Function. If the next-to-execute Pallet is not appointed, the executionright is transferred to all Tense Control Vectors belonging to the Pallet Function, and the Tense Control Vectors implement re-execution. If the re-execution of Tense Control Vectors continues, the W0 s state

7 transition will disappear. The transfer of the executionright from W0 to W0 is a role of the Tense Control Function. Vesting the execution-right in the Tense Control Vectors is a role of the Pallet Function, whereas returning the execution-right to the Pallet Function is a role of Tense Control Vectors. Vesting the executionright in the Pallet Functions is a role of the Tense Control Function, whereas returning the execution-right to the Tense Control Function is a role of the Pallet Functions. R is the Routing Vector of the W0 Pallet. The st rule of R checks the state transition of the W0 to which the R belongs. If there is no state transition, the R s th rule appoints W0 of the same F into the R s th area as a Pallet to be executed next, and the execution-right is returned to the Pallet Function. If there is any state transition, nothing is executed, and the execution-right is returned to the Pallet Function. RR, RD, RM and RC are the Routing Vectors of W0. The st rule of RR checks the state transition of the W0 to which the RR belongs. If there is any state transition, W0 of the same F is appointed into the RR s th area as the next-to-execute Pallet. If there is no state transition, and the execution-right is returned to the Pallet Function. The st rule of RD checks the state transition of the W0 to which the RD belongs. If there is no state transition, W0 of the F that is located just before this operating F on PRD is appointed into the RD s th area as the Pallet to be executed next. If there is no state transition, the execution-right is returned to the Pallet Function. While there is state transition in W0, RR handles the situation. When termination of the operation of F is not declared, the st rule of RC checks the state transition of the W0 to which the RC belongs. If there is no state transition, W0 of the succeeding F on PRD is appointed into the RC s th area as the Pallet to be executed next. If there is state transition, nothing is executed and the execution-right is returned to the Pallet Function. While there is state transition in W0, RR handles the situation. When termination of the operation of F is declared and if F to be executed next is specified on PRD, RC appoints W0 of the next-to-execute F into the RC s th area. Thus, RC are defined separately for the cases that termination of the operation of F is declared and not declared. If termination of the operation of F is declared, RM appoints into its own th area the W0 of the preceding F to return as the Pallet to be executed next. This preceding F is located two or more Fs before this operating F and is specified as the F to be executed next on PRD. 8. Nature of noun Meaning and functionality that cannot be realized with nouns extracted mechanically or with the complementary action can be realized by extending the concept of the noun. The extended concept of the nouns are called Boundary Noun, quivalent Noun and ummation Noun. Propositions are made for each of these nouns. 8.. Boundary noun When three Fs, which are denoted as F, F and F respectively, make a row connected with RD on PRD, F and F can be treated as one F, and F and F can also be treated as one F. However, it is uncertain that F and F possess a relation to materialize the ynchronous tructure. Therefore, if the th area of the ignification Vector of a certain noun of W0 of F is used for the complementary action of the ignification Vector of W0 of F, there is a case in which the ynchronous tructure of F does not materialize. In order to avoid this concern, the th area of the noun of the W0 ignification Vector of F is duplicated and placed also in the W0 area of F. When the th area of a noun on W0 is justified, its ignification Vector sets the result not only into its th area but also into the th area of the same noun duplicated in the F. uch word is called Boundary Noun. Whether a Boundary Noun is required or not is determined by the nd rule of the ignification Vector of W0 of F. 8.. quivalent noun The purpose of W0 ignification Vector is to determine output data. The way of the generation of the data happens to become plural depending upon the condition specified by W0 ignification Vector. In this instance, one W0 ignification Vector for the noun is defined, whereas W0 ignification Vectors for the same noun for each way of the data generation should be defined separately. The noun having plural W0 ignification Vectors is called quivalent Noun. 8.. ummation noun There are nouns to be used for arithmetic summation. An auxiliary memory area for the exclusive use for summing up is added in the nd area of the W0 ignification Vectors of those nouns.

8 9. Database In the conventional ways of software development, the definition of data items of database is implemented logically, so there is a fundamental problem that logical adjustment between the database and the programs cannot be easily done. F is comprosed of ignification Vectors. ignification Vectors are determined for each noun. ignification Vectors are independent to each other. This gives universality to the logical relation between database and programs. Consequently, design work of database becomes easy. Proceedings of the ACI nd International Conference on oftware ngineering, Artificial Intelligence, Networking & Parallel/Distributed Computing (NPD 0). August 00. Nagoya, Japan. 0. Conclusion The purpose of this paper is to present a summary of the scheme of TDM and the structure of F to give a comprehensive view of the theory, and the some detailed information necessary for actual implementation is omitted for his purpose. The detailed information is shown at our web site. F is created based on a non-engineering hypothesis. However, in the actual software development, high productivity and maintainability are witnessed. This is the first case that non-engineering hypothesis solves issues arising in the world of software engineering. The program structure of F is completely different from that of conventional programs, which is an inevitable conclusion. The concept that the F s structure shows is new and different from the engineering concept of software. For example, the pre-execution and proexecution states are the same in the conventional program structure, whereas F is different. The difference brings advantages to software developers. If the environments to determine F, such as tools, O and programming languages, are developed, the level of world of software would surpass the present one. The hypothesis presented here is deeply influenced with the works of pinoza, Leibniz and Wittgenstein but not based on engineering. This is why the author uses the term non-engineering. This study is based on theory established in the Consciousness Model, and the theory will be presented at another opportunity. RFRNC. F. Negoro, Principle of Lyee oftware, in Proc. I000, Aizu, Japan, Nov. 000, pp. -.. F. Negoro, A Proposal for Requirement ngineering, to be included in Proc. ADBI 00, Vilnius, Lithuania, ep F. Negoro, Intent Operationaliasation for ource Code Generation, to be included in the Proc. CI00, Orlando, UA, Jul